Information processing system, information processing method, and program

Abstract

To provide a technology that enables appropriate searching of data stored in a predetermined database. [Solution] An information processing system comprising: a reception unit that receives a sentence input by a user; a first extraction unit that extracts one or more keywords from the sentence; a calculation unit that calculates the frequency of appearance of at least the one or more keywords in the sentence; a second extraction unit that extracts one or more specific keywords from the one or more keywords based on the frequency of appearance; and a first generation unit that generates a query for searching data stored in a specified database based on the sentence and the one or more specific keywords.

Description

[Technical Field]

[0001] The present invention relates to an information processing system, an information processing method, and a program. [Background technology]

[0002] 2. Description of the Related Art Conventionally, techniques capable of automatically generating graphs and the like from predetermined data using a computer are widely known.

[0003] In this regard, Patent Document 1 describes a display system in which a user inputs the purpose for creating a graph (analysis purpose), extracts keywords from the sentence entered by the user indicating the analysis purpose, and generates a graph related to the keywords from specified data based on the keywords. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 2020-003880 Summary of the Invention [Problem to be solved by the invention]

[0005] However, with the above-described conventional techniques, it is not possible to generate a query that more accurately reflects the user's intentions, and there is a risk that data stored in a predetermined database cannot be searched appropriately.

[0006] The present invention has been made in view of the above, and has an object to provide a technique that enables appropriate searching of data stored in a predetermined database. [Means for solving the problem]

[0007] An information processing system according to one embodiment of the present invention includes a receiving unit that receives a sentence input by a user, a first extraction unit that extracts one or more keywords from the sentence, a calculation unit that calculates the frequency of occurrence of at least the one or more keywords in the sentence, a second extraction unit that extracts one or more specific keywords from the one or more keywords based on the frequency of occurrence, and a first generation unit that generates a query for searching data stored in a specified database based on the sentence and the one or more specific keywords.

[0008] An information processing method according to one aspect of the present invention is an information processing method executed by a computer, and includes the steps of accepting a sentence input by a user, extracting one or more keywords from the sentence, calculating the frequency of occurrence of at least the one or more keywords in the sentence, extracting one or more specific keywords from the one or more keywords based on the frequency of occurrence, and generating a query for searching data stored in a specified database based on the sentence and the one or more specific keywords.

[0009] A program according to one embodiment of the present invention causes a computer to function as a reception unit that receives a sentence input by a user, a first extraction unit that extracts one or more keywords from the sentence, a calculation unit that calculates the frequency of occurrence of at least the one or more keywords in the sentence, a second extraction unit that extracts one or more specific keywords from the one or more keywords based on the frequency of occurrence, and a first generation unit that generates a query for searching data stored in a specified database based on the sentence and the one or more specific keywords. [Effects of the Invention]

[0010] According to the aspects of the present invention, data stored in a predetermined database can be appropriately searched. [Brief explanation of the drawings]

[0011] [Figure 1]1 is a diagram illustrating an overall configuration of an image generation system according to an embodiment. [Figure 2] FIG. 10 is a diagram illustrating an example of imported information according to an embodiment. [Figure 3] FIG. 10 is a diagram illustrating an example of schema information according to an embodiment. [Figure 4] FIG. 10 is a diagram illustrating an example of a code according to an embodiment. [Figure 5] FIG. 10 is a diagram illustrating another example of a code according to an embodiment. [Figure 6] FIG. 10 is a diagram illustrating an example of data used in a dictionary generation process according to an embodiment. [Figure 7] FIG. 10 is a diagram illustrating an example of a dictionary according to an embodiment; [Figure 8] FIG. 10 is a diagram illustrating an example of code related to a dictionary generation process according to an embodiment. [Figure 9] 10 is a flowchart illustrating an example of an image generation process according to an embodiment. [Figure 10] FIG. 10 is a diagram illustrating an example of code related to an image generation process according to an embodiment. [Figure 11] FIG. 10 is a diagram illustrating an example of an image generated by an image generation process according to an embodiment. [Figure 12] FIG. 10 is a diagram illustrating another example of an image generated by the image generation process according to an embodiment. [Figure 13] FIG. 10 is a diagram illustrating another example of an image generated by the image generation process according to an embodiment. [Figure 14] FIG. 10 is a diagram illustrating another example of an image generated by the image generation process according to an embodiment. [Figure 15] FIG. 10 is a diagram illustrating another example of an image generated by the image generation process according to an embodiment. [Figure 16] 10 is a flowchart illustrating an example of a moving image generation process according to an embodiment. [Figure 17] FIG. 2 is a diagram illustrating an example of a hardware configuration of a computer according to an embodiment.

[0012] Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the embodiments described below are merely examples, and are not intended to exclude various modifications or applications of techniques not explicitly described below. In other words, the present invention can be implemented in various modifications (such as combining the respective embodiments) without departing from the spirit of the invention. Furthermore, in the following description of the drawings, identical or similar parts are denoted by the same or similar reference numerals.

[0013] In the embodiments of the present invention, the terms "unit," "device," and "system" do not simply mean physical means, but also include cases where the functions of the "unit," "device," and "system" are realized by software. Furthermore, the functions of one "unit," "device," and "system" may be realized by two or more physical means or devices, and the functions of two or more "units," "devices," and "systems" may be realized by one physical means or device.

[0014] 1 is a diagram showing the overall configuration and overview of an image generation system (information processing system) according to one embodiment. The image generation system 100 is an image generation system that can, for example, generate a query that more accurately reflects a user's intention and appropriately search data stored in a predetermined database in order to generate a predetermined image. The image generation system 100 illustratively includes an image generation server 1 that generates a predetermined image, and a user terminal 3 operated by a user U.

[0015] Here, an overview of an image generation system 100 according to one embodiment of the present invention will be described. In conventional data analysis and search, unless a user anticipates the output data (or understands in advance the contents of all data of the target specifications) and then inputs search keywords, there is a risk that the desired data cannot be extracted or data that does not meet the user's intentions may be extracted.

[0016] To solve this problem, the image generation system 100 accepts a prompt (question sentence) input by a user and extracts important keywords from the prompt and all data of the specifications. The image generation system 100 predicts the user's intention based on the extracted important keywords, corrects or updates the prompt based on the predicted intention using a large language model (LLM), and provides analysis and search results that are in line with the user's intention. Note that LLMs are learning models that learn from large amounts of text data and perform natural language processing tasks. Examples of large language models include GPT-4, BERT, LaMDA, and PaLM.

[0017] For example, in the case of a normal search, if a user enters the prompt "ramen," a variety of information is extracted, such as popular ramen, types of ramen, ramen near train stations, ramen recipes, and the history of ramen. From this, various search results can be output according to the user's prompt. In such cases, it is highly likely that search results that do not match (are off from) the user's intentions will be output. To address this problem, when a user inputs a prompt, the image generation system 100 grasps the user's intentions, such as which keywords are important in all the specification data, and what the user is looking for based on the overlap of one or more important keywords, and then outputs the data search results.

[0018] More specifically, the image generation system 100 includes a receiving unit that receives a sentence input by a user, a first extraction unit that extracts one or more keywords from the sentence, a calculation unit that calculates the appearance frequency of at least the one or more keywords in the sentence, a second extraction unit that extracts one or more specific keywords from the one or more keywords based on the appearance frequency, and a first generation unit that generates a query for searching data stored in a predetermined database based on the sentence and the specific one or more keywords. This image generation system 100 can generate a query that more accurately reflects the user's intention. Therefore, data stored in the predetermined database can be appropriately searched. This image generation system 100 can then generate an image desired by the user based on the searched data.

[0019] In the image generation system 100 shown in FIG. 1, the image generation server 1 and the user terminal 3 are connected to each other via a predetermined communication network. The predetermined communication network may be a wired communication network or a wireless communication network. The image generation server 1 is configured, for example, by a server device configured by a computer system or a cloud server, and is implemented by a software program. The user terminal 3 may be, for example, a general-purpose personal computer, a smartphone, a tablet terminal, or other device. The number of user terminals 3 is arbitrary, and the user terminal 3 is assumed to be one or more user terminals.

[0020] As shown in FIG. 1, the image generation server 1 functionally includes an information processing unit 10 that executes various processes for image generation, and a database (DB) 5 that records information for executing various processes for image generation and the execution results of the various processes. The DB 5 is a searchable database that stores all data (e.g., text data, still image data, and video image data) contained in data files related to various software, such as Excel files, Word files, and PDF files. The DB 5 records table information and field information (e.g., information related to table definitions and schemas), which will be described later. The DB 5 also records data generated by the feature calculation unit 13, which will be described later. If the data stored in the DB 5 includes at least one or more video data files, the format of the one or more video data files is vector data. However, the format of the one or more video data files is not limited to the above and may be other data formats.

[0021] As described above, the image generation server 1 is configured by, for example, a server device configured by a computer system, a cloud server, etc., and is implemented by a software program. The program may be provided by being downloaded from outside the image generation server 1 (for example, an external device or cloud) via a predetermined network, or may be provided by various computer-readable information recording media such as CD-ROM and DVD-ROM.

[0022] The information processing unit 10 functionally includes a file acquisition unit 11, a feature calculation unit 13, a text data division unit 15, an information storage unit 17, a dictionary generation unit 19, a prompt reception unit 21 (reception unit), a keyword division unit 23 (first extraction unit), a keyword extraction unit 25 (calculation unit and second extraction unit), a prompt completion unit 27, an information provision unit 29, an SQL query generation unit 31 (first generation unit), a text display control unit 33, a text display unit 35, a strategy analysis unit 37 (evaluation unit), an image display control unit 39 (second generation unit), and an image display unit 41. Note that SQL (Structured Query Language) is one of computer languages ​​and includes, for example, a language for manipulating data stored in a database.

[0023] The file acquisition unit 11 acquires data files related to various software, such as Excel files, Word files, and PDF files.

[0024] The feature calculation unit 13, for example, cleanses, converts, and aggregates (feature processing) each acquired file. The information storage unit 17 stores the data processed by the feature calculation unit 13 in the DB 5. The information storage unit 17 stores, for example, field information (e.g., information related to table definitions and schemas) in the DB 5.

[0025] An example of information storage processing in DB5 will be described with reference to FIGS. 2 to 5. FIG. 2 is a diagram showing an example of imported information according to an embodiment. FIG. 3 is a diagram showing an example of schema information according to an embodiment. FIGS. 4 and 5 are diagrams showing an example of code according to an embodiment. The information storage unit 17, for example, identifies the data type of each field in the data file, and stores the identified information as schema information (for example, field configuration information) in DB5. As shown in FIGS. 2 and 3, for example, the information storage unit 17 stores the contents of each field as "detail" while summarizing them using, for example, LLM.

[0026] 4, for example, the dtypes of Pandas (a Python data analysis library that provides functions to facilitate data analysis) are actually managed as two-dimensional arrays using NumPy, and field values ​​of string type, list type, etc. can be saved as object type. For example, if a field is of string type, the information storage unit 17 can check the length of all records of this field and infer the data type.

[0027] As shown in FIG. 5, the information storage unit 17 creates a table and stores data using a CREATE or INSERT statement based on the above data. This makes it possible to completely eliminate file upload requests based on the format, structure, and field configuration predetermined by the system designer. Note that while Python and Pandas are used as examples in this embodiment, other programming languages ​​and libraries can be used. For example, Polars and Pandas can be used in this embodiment.

[0028] An example of dictionary generation processing will be described with reference to FIGS. 6 to 8. FIG. 6 is a diagram showing an example of data used in dictionary generation processing according to an embodiment. FIG. 7 is a diagram showing an example of a dictionary according to an embodiment. FIG. 8 is a diagram showing an example of code related to dictionary generation processing according to an embodiment. As shown in FIGS. 6 to 8, the text data division unit 15 shown in FIG. 1 divides text data included in each acquired file using a predetermined natural language process. The text data division unit 15 divides character string data into words (sentences) using morphological analysis, for example, using at least one of TF-IDF and BM25. The dictionary generation unit 19 shown in FIG. 1 generates a dictionary (list) based on the text data (e.g., keywords) divided by the text data division unit 15, and stores the list in DB5.

[0029] However, users may not always know the field and data format (e.g., string and numeric) in which their files are stored in the database, making it difficult to identify the desired data. Therefore, even if users enter a query using as many general keywords as they can think of as a prompt, there is a concern that the SQL query generated through LLM will not yield the desired search results. Furthermore, if the entered keywords are simple words such as "update," "type," "store," and "food," a large number of search results may be obtained, but it may be difficult to find the desired information. Therefore, by generating a list like the one described above, users can easily access the desired data.

[0030] The prompt receiving unit 21 (receiving unit) receives a prompt (sentence) input by a user. The keyword dividing unit 23 (first extracting unit) divides and extracts one or more keywords from the prompt received by the prompt receiving unit 21. The keyword dividing unit 23 divides and processes the text data included in the received prompt using predetermined natural language processing.

[0031] The keyword extraction unit 25 (calculation unit) calculates at least the frequency of appearance of one or more keywords in the received prompt. The keyword extraction unit 25 (second extraction unit) extracts one or more specific keywords from the one or more keywords based on the calculated frequency of appearance. The keyword extraction unit 25 calculates the frequency of appearance of one or more keywords using at least one of TF-IDF and BM25 processing, for example, and extracts the specific one or more keywords.

[0032] The keyword extraction unit 25 may calculate the frequency of appearance of one or more keywords in the prompt and DB5 (a predetermined database), and extract one or more specific keywords from the one or more keywords based on the frequency of appearance. With this configuration, the keyword extraction unit 25 can calculate the frequency of appearance by further referring to one or more keywords in DB5 in addition to the prompt. Since prompts from users often contain terms that are not commonly used in DB5, selecting keywords by taking into consideration, for example, expressions that appear frequently in DB5 can improve the accuracy of the SQL query and increase the probability of obtaining the desired results.

[0033] The keyword extraction unit 25 generates a list in which one or more superordinate keywords and one or more subordinate keywords are associated with each other based on one or more specific keywords. For example, as shown in the example of Fig. 7, the keyword extraction unit 25 can generate a list in which the superordinate keyword "price" is associated with the subordinate keywords "650 yen" and "50 yen." Note that while the example shown in Fig. 7 is an example of list creation based on an acquired data file, the example described here is an example of list creation based on text data corresponding to a prompt entered by the user.

[0034] The prompt completion unit 27 completes the prompt based on the keywords divided by the keyword division unit 23 and the list generated by the keyword extraction unit 25. The information provision unit 29 provides the SQL query generation unit 31 with table information (e.g., table data and field information) and schema information recorded in the DB 5.

[0035] The SQL query generation unit 31 (first generation unit) generates an SQL query (query) for searching data stored in DB5 (a specified database) based on the prompt received by the prompt reception unit 21 and one or more specific keywords extracted by the keyword extraction unit 25.

[0036] Here, if a user were to search for data using the keyword "ramen" in a given database, they might enter a prompt like, "Where can I find a good ramen restaurant?" However, if the list referenced in the data search is not specialized in a specific theme, information such as "noodle dishes" and "tsukemen" may not be rated. Furthermore, it is unclear whether the original list classifies and rates only a small subset of information about "ramen" from various categories of restaurant reviews (a bird's-eye view), or whether it rates each review from a list of all ramen restaurants (a micro-view). As such, it is difficult for both the user and the system to understand the nature of the list and then identify and classify only the review fields.

[0037] Therefore, when a user enters a prompt related to, for example, "ramen," the SQL query generation unit 31 can divide the prompt using BM25 or the like and include, via a corpus, related information such as "noodle dishes" and "tsukemen." With this configuration, the SQL query generation unit 31 can generate information that facilitates an answer using an SQL query while identifying fields that require an answer to the prompt based on table definitions and schema information using LLM. Based on the user's prompt (e.g., a prompt completed by the prompt completion unit 27) and the table and schema information, the SQL query generation unit 31 generates an SQL query that prompts an appropriate answer from DB5 using LLM. The SQL query generation unit 31 may update the SQL query based on the evaluation results of the strategy analysis unit 37.

[0038] The text display control unit 33 generates output information based on an SQL query that includes information about the output format of the output information for outputting text to the user terminal 3 shown in Fig. 1. Upon receiving the output information generated by the text display control unit 33, the text display unit 35 outputs text to the user terminal 3 based on the output information.

[0039] The strategy analysis unit 37 (evaluation unit) evaluates the retrieved data based on the SQL query generated by the SQL query generation unit 31 and the data retrieved based on the SQL query. The evaluation method will be described later.

[0040] The image display control unit 39 (second generation unit) generates output information based on an SQL query including information regarding the output format of the output information for outputting an image to the user terminal 3. When the image display unit 41 receives the output information generated by the image display control unit 39, it outputs an image to the user terminal 3 based on the output information. Note that the information regarding the output format of the output information for outputting an image may include setting information regarding the type of graph, chart, and / or dashboard, legend items, axis labels, and / or scale.

[0041] <Image generation processing> An example of image generation processing according to an embodiment will be described with reference to FIGS. 9 to 15. FIG. 9 is a flowchart showing an example of image generation processing according to an embodiment. As a premise, in the image generation processing, image generation application software including the image generation processing according to an embodiment of the present invention is downloaded from, for example, a predetermined site on a network or a recording medium, and stored in an executable manner in the image generation server 1 shown in FIG. 1. Then, when an instruction to execute the image generation processing is given, program operation based on the software starts. Note that the software is also executed in advance in the user terminal 3 shown in FIG. 1 so that it operates appropriately as a component of the image generation system 100. Furthermore, the same prerequisite processing as the image generation processing is executed for the video generation processing described below.

[0042] As shown in Fig. 9, the image generation server 1 shown in Fig. 1 accepts a prompt input by a user (step S1). The image generation server 1 determines whether chart generation is necessary based on the input prompt (step S3). If it is determined that chart generation is unnecessary (No), the chart generation process ends. On the other hand, if it is determined that chart generation is necessary (Yes), the process proceeds to step S5.

[0043] The image generation server 1 derives a data analysis method (step S5). The image generation server 1 (strategy analysis unit 37 shown in FIG. 1) analyzes a strategy for chart generation using LLM based on the SQL query results (e.g., records and values) and the user prompts (e.g., prompts modified as necessary). In this strategy analysis, the image generation server 1 can derive an appropriate data analysis method (e.g., at least one of bar graphs, line graphs, ABC analysis, basket analysis, budget-actual analysis, and defect rate analysis).

[0044] The image generation server 1 evaluates the retrieved data based on the generated SQL query and the data retrieved based on the SQL query (step S7). For example, the image generation server 1 checks whether the necessary SQL query results are available. More specifically, it evaluates whether the data type obtained from the SQL query covers the information necessary for the data analysis method (e.g., chart type) recommended by the LLM.

[0045] For example, when attempting to graph hourly precipitation data for a certain city, if the results retrieved by the SQL query from DB5 shown in Figure 1 show precipitation data for every three hours, or precipitation data for a different city, it is determined that the query results are incomplete (insufficient). In this case (No in step S7), the process proceeds to step S9. Based on the result of step S7, the image generation server 1 generates additional SQL queries to obtain the necessary values ​​and records in order to obtain sufficient query results (step S9).

[0046] On the other hand, if the query results are all available (Yes in step S7), the process proceeds to step S11. The image generation server 1 generates the code required to generate the chart (step S11).

[0047] 10 is a diagram showing an example of code related to an image generation process according to an embodiment. As shown in FIG. 10, the image generation server 1 provides code necessary for chart generation, such as specifying the chart type, legend items, and horizontal axis labels in response to user prompts. This code can use, for example, a web framework called Svelte, to render the code necessary for web page and chart generation on a browser.

[0048] The image generation server 1 executes a validation check (e.g., confirmation and verification of validity) of the generated code (step S13). If the validation requirements are met (Yes in step S13), the process proceeds to step S15, and if the validation requirements are not met (No in step S13), the process returns to step S11.

[0049] The image generation server 1 generates a chart based on the generated code (step S15). Figures 11 to 15 are diagrams showing examples of images generated by the image generation process. As shown in Figures 11 to 15, the image generation server 1 generates charts including bar graphs, line graphs, etc. that meet the user's intentions based on prompts input by the user, and displays them on the user terminal 3 shown in Figure 1.

[0050] According to the image generation process described above, the image generation system 100 can generate an SQL query that more accurately reflects the user's intentions when generating an image. Therefore, it is possible to appropriately search for data for image generation stored in a predetermined database. Then, the image generation system 100 can generate an image desired by the user based on the searched data.

[0051] Although an example of a process for generating an image including a graph, a chart, and / or a dashboard has been described above as one embodiment of the present invention with reference to Figures 9 to 15, the present invention is not limited to this. The present invention can also be applied to a moving image generation process as one embodiment.

[0052] <Video generation processing> An example of video generation processing according to an embodiment will be described with reference to FIG. 16. FIG. 16 is a flowchart showing an example of video generation processing according to an embodiment. As shown in FIG. 16, the image generation server 1 shown in FIG. 1 uploads video data based on a user instruction (step S21). The image generation server 1 converts the uploaded video data into vector data and stores it in the DB 5 shown in FIG. 1 (step S22). The image generation server 1 uses a predetermined model to detect objects of vectors on the x, y, and z axes from the camera coordinates in the video data, adds information about the time period, and stores the data in the DB 5 (step S23).

[0053] The image generation server 1 receives a prompt from the user (step S24). The image generation server 1 generates an SQL query based on the prompt and table information recorded in DB5, for example, using LLM (step S25). The image generation server 1 searches for data for generating a moving image recorded in DB5 based on the SQL query (step S26). The image generation server 1 generates and outputs a moving image based on the searched data (step S27).

[0054] According to the above-described video generation process, the image generation system 100 can generate an SQL query that more accurately reflects the user's intentions when generating a video. Therefore, the image generation system 100 can appropriately search for data for generating a video stored in a predetermined database. The image generation system 100 can then generate a video desired by the user based on the searched data.

[0055] Fig. 17 is a diagram showing an example of the hardware configuration of a computer according to an embodiment of the present invention. With reference to Fig. 17, an example of the hardware configuration of a computer that can be used to configure the image generation server 1 and the user terminal 3 shown in Fig. 1 will be described.

[0056] 17, a computer 200 mainly comprises, as hardware resources, a processor 201, a main storage device 202, an auxiliary storage device 203, an input / output interface 204, and a communication interface 205, which are interconnected via a bus line 206 including an address bus, a data bus, a control bus, etc. Note that an interface circuit (not shown) may be interposed between the bus line 206 and each hardware resource as appropriate.

[0057] The processor 201 controls the entire computer. The processor 201 corresponds to, for example, the information processing unit 10 shown in FIG. 2. The main storage device 202 provides a working area for the processor 201 and stores an SRAM ( S tatic R andom A access M emory) and DRAM ( D Dynamic R andom A access M The auxiliary storage device 203 is a volatile memory such as a memory card (RAM) or a hard disk drive (HDD), and is a non-volatile memory such as a hard disk drive (HDD), solid state drive (SSD), or flash memory that stores software programs and data. The programs and data are loaded from the auxiliary storage device 203 to the main storage device 202 via the bus line 206 at any time.

[0058] The input / output interface 204 presents information and / or receives input of information, and may be a camera, keyboard, mouse, display, touch panel display, microphone, speaker, etc. The communication interface 205 is connected to the above-mentioned predetermined communication network and transmits and receives data via the predetermined communication network. The communication interface 205 may also acquire information related to the network, such as information related to Wi-Fi access points.

[0059] It will be apparent to those skilled in the art that the cooperation of the hardware resources and software exemplified above enables the computer 200 to function as desired means, execute desired steps, and achieve desired functions.

[0060] It should be noted that the above-described embodiments are provided to facilitate understanding of the present invention and are not to be construed as limiting the present invention. The present invention may be modified or improved without departing from the spirit thereof, and equivalents thereof are also included in the present invention. Furthermore, the present invention can be formed into various disclosures by appropriately combining multiple components disclosed in the above-described embodiments. For example, some components may be deleted from all the components shown in the embodiments. Furthermore, components may be appropriately combined in different embodiments. [Explanation of symbols]

[0061] 1...image generation server, 3...user terminal, 5...database (DB), 10...information processing unit, 11...file acquisition unit, 13...feature calculation unit, 15...text data division unit, 17...information storage unit, 19...dictionary generation unit, 21...prompt reception unit, 23...keyword division unit, 25...keyword extraction unit, 27...prompt completion unit, 29...information provision unit, 31...SQL query generation unit, 33...text display control unit, 35...text display unit, 37...strategy analysis unit, 39...image display control unit, 41...image display unit, 100...image generation system, 200...computer, 201...processor, 202...main recording device, 203...auxiliary recording device, 204...input / output interface, 205...communication interface, 206...bus line,

Claims

1. a receiving unit that receives a sentence input by a user; a first extraction unit that extracts one or more keywords from the sentence; a calculation unit that calculates the frequency of appearance of the one or more keywords in at least the sentence; a second extraction unit that extracts one or more specific keywords from the one or more keywords based on the appearance frequency; a first generation unit that generates a query for searching data stored in a predetermined database based on the sentence and the one or more specific keywords; Information processing system.

2. the query includes information regarding an output format of output information for outputting an image to a user terminal operated by the user, a second generation unit that generates the output information based on the query; The information processing system according to claim 1 .

3. the calculation unit calculates the frequency of appearance of the one or more keywords in the sentence and the predetermined database; the second extraction unit extracts one or more specific keywords from the one or more keywords based on the appearance frequency; The information processing system according to claim 1 .

4. the second extraction unit generates a list in which one or more keywords of a higher concept are associated with one or more keywords of a lower concept, based on the one or more specific keywords; The information processing system according to claim 1 .

5. an evaluation unit that evaluates the retrieved data based on the query and the data retrieved based on the query; the first generation unit updates the query based on the evaluation result of the evaluation unit. The information processing system according to claim 1 .

6. The information regarding the output form of the output information includes setting information regarding the type of graph, chart, and / or dashboard, legend items, axis labels, and / or scales; The information processing system according to claim 2 .

7. The data stored in the predetermined database includes at least one or more video data files, The one or more video data files are in a vector data format. The information processing system according to claim 1 .

8. 1. A computer-implemented information processing method, comprising: accepting a sentence input by a user; extracting one or more keywords from the text; calculating the frequency of appearance of the one or more keywords in at least the sentence; extracting one or more specific keywords from the one or more keywords based on the frequency of appearance; generating a query for searching data stored in a predetermined database based on the sentence and the one or more specific keywords; Information processing methods.

9. Computer, a receiving unit that receives a sentence input by a user; a first extraction unit that extracts one or more keywords from the sentence; a calculation unit that calculates the frequency of appearance of the one or more keywords in at least the sentence; a second extraction unit that extracts one or more specific keywords from the one or more keywords based on the appearance frequency; a first generation unit that generates a query for searching data stored in a predetermined database based on the sentence and the one or more specific keywords; A program to function as a

Images